Regulation of CD38 expression in human airway smooth muscle cells: role of class I phosphatidylinositol 3 kinases

Abstract

The ADP-ribosyl cyclase activity of CD38 generates cyclic ADP-ribose, a Ca(2+)-mobilizing agent. In human airway smooth muscle (HASM) cells, TNF-α mediates CD38 expression through mitogen-activated protein kinases and NF-κB and AP-1. The phosphatidylinositol-3 kinase/Akt (PI3K/Akt) pathway is involved in TNF-α signaling and contributes to airway hyperresponsiveness and airway remodeling. We hypothesized that PI3Ks mediate CD38 expression and are involved in the differential induction of CD38 by TNF-α in asthmatic HASM cells. HASM cells were treated with pan-PI3K inhibitors (LY294002 or wortmannin) or class I-selective (GDC0941) or isoform-selective PI3K inhibitors (p110α-PIK-75 and p110β-TGX-221) with or without TNF-α. HASM cells were transfected with a catalytically active form of PI3K or phosphatase and tensin homolog (PTEN) or nontargeting or p110 isoform-targeting siRNAs before TNF-α exposure. CD38 expression and activation of Akt, NF-κB, and AP-1 were determined. LY294002 and wortmannin inhibited TNF-α-induced Akt activation, whereas only LY294002 inhibited CD38 expression. P110 expression caused Akt activation and basal and TNF-α-induced CD38 expression, whereas PTEN expression attenuated Akt activation and CD38 expression. Expression levels of p110 isoforms α, β, and δ were comparable in nonasthmatic and asthmatic HASM cells. Silencing of p110α or -δ, but not p110β, resulted in comparable attenuation of TNF-α-induced CD38 expression in asthmatic and nonasthmatic cells. NF-κB and AP-1 activation were unaltered by the PI3K inhibitors. In HASM cells, regulation of CD38 expression occurs by specific class I PI3K isoforms, independent of NF-κB or AP-1 activation, and PI3K signaling may not be involved in the differential elevation of CD38 in asthmatic HASM cells.

Figure 1.

Effects of pan-PI3 kinase inhibitors on TNF-α–induced CD38 expression in human airway smooth muscle (HASM) cells. (A) Growth-arrested HASM cells were treated with TNF-α for variable lengths of time (0–120 min). Note the time-dependent increase in TNF-α–induced Akt activation (left panel). In the presence of LY294002 (3 μM) or wortmannin (100 nM), there was no detectable activation of Akt in response to 0 or 30 minutes of exposure to TNF-α (right panel) (blot representative of four independent experiments). In the presence of LY294002 or wortmannin, exposure to TNF-α for 120 minutes failed to induce Akt phosphorylation (data not shown). (B) CD38 mRNA expression in the presence of PI3 kinase inhibitors. Representative agarose gel (upper panel) and quantitative RT-PCR (lower panel) show that TNF-α–induced CD38 mRNA expression is attenuated by LY294002 (LY+T) with no apparent inhibitory effect by wortmannin (W+T) (average of six independent experiments). (C) ADP-ribosyl cyclase activity was determined after TNF-α exposure (TNF) in the presence or absence of LY294002 or wortmannin. Note the inhibition of TNF-α–induced ADP-ribosyl cyclase activity in the presence of LY294002 (T + LY), with no significant change in ADP-ribosyl cyclase activity in the presence of wortmannin (T + Wort) (average of six independent experiments). P ≤ 0.05. a = significant compared with vehicle control; b = significant compared with the TNF-α treatment).

Figure 2.

Effects of transient expression of PI3 kinase and phosphatase and tensin homolog (PTEN) on Akt activation and CD38 expression. (A) Control vector (pSG5) or vectors carrying PI3 kinase catalytic subunit (p110) or PTEN construct were transfected into HASM cells and treated with TNF-α, and the lysates were immune blotted for phosphorylated (pThr³⁰⁸) and total Akt and p-PTEN. Note the elevated basal and TNF-α–induced Akt activation in p110-transfected cells (lanes 5 and 6) compared with the control vector–transfected cells (lanes 3 and 4). PTEN transfection decreased the basal and TNF-α–induced Akt activation (lanes 7 and 8) (blot representative of three independent experiments). (B) Bar graph shows the increase in p-Akt band intensity relative to the vector-transfected controls (average of three independent experiments). (C) TNF-α–induced CD38 expression in cells after transient transfection with p110 or PTEN. Representative gel image shows CD38 mRNA expression in the transfected cells. Lanes 3, 5, 9, and 11: CD38 expression in cells transfected with vector and treated with vehicle. Lanes 4, 6, 10, and 12: CD38 expression in transfected cells after exposure to TNF-α. Note the elevated basal (lane 5) and TNF-α–induced (lane 6) CD38 expression in p110-transfected cells. There was a lack of significant effect of PTEN transfection on basal CD38 expression (lane 11), and TNF-α–exposure elicited only a slight increase of CD38 expression (lane 12). The image is representative of three independent experiments.

Figure 3.

Expression of p110 isoforms in nonasthmatic and asthmatic HASM cells. Total cell lysates obtained from nonasthmatic and asthmatic HASM cells were immune blotted to determine expression of class IA p110 isoforms α, β, and δ. (A) Expression of isoforms α, β, and δ were comparable in HASM cells obtained from donors with and without asthma (blots representative of four independent experiments). Densitometric analysis of the blots (n = 4) showing comparable expression of p110 isoforms α, β, and δ between HASM cells obtained from nonasthmatic (NA) and asthmatic (A) donors. (B) The p110 isoform γ, which is categorized as the member of class IB PI3 kinases, was not detected in HASM cells. The anti-p110γ antibody detected the p110γ isoform in Jurkat cell lysate (blot representative of four independent experiments).

Figure 4.

Effects of class I– or isoform-selective PI3 kinase inhibitors on TNF-α–induced CD38 expression in HASM cells. The HASM cells were treated with vehicle or TNF-α in the presence of GDC0941 (class I PI3 kinase-selective inhibitor), PIK-75 (p110α-selective inhibitor), TGX-221 (p110β-selective inhibitor), or the nonisoform selective PI3 kinase inhibitors LY294002 or wortmannin. (A) Representative blot showing the inhibitory effects of the PI3 kinase inhibitors on TNF-α–induced Akt activation (blot representative of three independent experiments). (B) Densitometry ratio of p-Akt/total Akt band intensity in cells exposed to the class I– or isoform-selective PI3 kinase inhibitors (average of three independent experiments). Note the partial inhibition of Akt activation by the class-I– or p110 isoform-selective inhibitors compared with the pan-PI3 kinase inhibitor wortmannin (*P < 0.05 compared with TNF-α treatment). (C) TNF-α–induced CD38 mRNA expression in cells exposed to PI3 kinase inhibitors. The class I– and isoform-selective inhibitors partially inhibited the CD38 mRNA expression, although the reduction did not reach statistical significance. The pan-PI3 kinase inhibitor LY294002 significantly inhibited the CD38 mRNA expression (average of three independent experiments; *P < 0.05 compared with TNF-α treatment). (D) ADP-ribosyl cyclase activity of CD38 was not altered by the class I–selective inhibitor (GDC0941) or the p110β-selective inhibitor (TGX-221). The p110α-selective inhibitor (PIK-75) and pan-PI3 kinase inhibitor LY294002 caused significantly decreased ADP-ribosyl cyclase activity (average of three independent experiments; *P < 0.05 compared with TNF-α treatment). (E) TNF-α induced differentially elevated CD38 mRNA expression in AASM cells compared with NAASM. TNF-α–induced CD38 expression in NAASM and AASM cells were comparably sensitive to the class I or isoform-selective PI3 kinase inhibitors (average of three independent experiments).

Figure 5.

Effects of p110 isoform-specific siRNA on CD38 mRNA expression in HASM cells. HASM cells were transfected with siRNA targeting p110α, -β, and -δ isoforms and exposed to vehicle or TNF-α 72 hours after transfection. (A) Representative blot showing down-regulation of each p110 isoform compared with scramble siRNA-transfected cells (representative of three independent experiments). (B) Bar graph showing significant down-regulation of each p110 isoform after transfection with relevant siRNA (average of three independent experiments). (C) In HASM cells transfected with p110 isoform–targeting siRNA, there were no apparent reductions in the basal or TNF-α–induced Akt activation (blot representative of six independent experiments; n = 3 for each NAASM and AASM group). (D) Bar graph showing densitometric analysis of p-Akt levels in NAASM cells transfected with p110 isoform–targeting siRNA (n = 3). (E) Bar graph showing densitometric analysis of p-Akt levels in AASM cells transfected with p110 isoform–targeting siRNA (n = 3). (F) Silencing of p110α or -δ isoform significantly attenuated TNF-α–induced CD38 mRNA expression, whereas silencing p110β did not have a significant effect on CD38 mRNA expression. The inhibitory effects of silencing p110 isoforms were comparable between NAASM and AASM cells (average of three independent experiments; *P < 0.05 compared with scrambled siRNA-transfected cells)

Figure 6.

Effects of PI3 kinase inhibitors on activation of transcription factors NF-κB and AP-1. HASM cells were treated with vehicle or TNF-α for 1 hour in the presence of class I isoform–selective or pan-PI3 kinase inhibitors, and the activation of transcription factors NF-κB or AP-1 was determined. (A) Representative blot shows that TNF-α–induced nuclear translocation NF-κB (p50 subunit) or AP-1 (p-c-Jun) was unaltered by the pan-PI3 kinase inhibitor LY294002. (B) TNF-α–induced NF-κB (i.e., p65 subunit binding to consensus sequence) activation was unaltered in the presence of class I– or isoform-selective PI3 kinase inhibitors or pan-PI3 kinase inhibitor LY294002. (C) TNF-α–induced AP-1 (i.e., p-c-Jun subunit binding to consensus sequence) activation was unaltered in the presence of class I– or isoform-selective PI3 kinase inhibitors or pan-PI3 kinase inhibitor LY294002.

Figure 7.

A proposed model for PI3 kinase regulation of CD38 expression in HASM cells. The model depicts signaling pathways known to regulate TNF-α–induced CD38 expression in HASM cells, according to findings from our laboratory and others. Induction of CD38 and other proinflammatory genes by TNF-α is mediated through TNFR1 (18, 20). Downstream of cytokine signaling, the small G-protein Ras is recruited and acts as an upstream regulator of PI3 kinase and MAPK signaling pathways (–23). The PI3 kinase converts phosphatidylinositol-3,4- bisphosphate (PIP₂) into phosphatidylinositol-3,4,5-trisphosphate (PIP₃). PIP₃ recruits pleckstrin homology domain-containing proteins, such as Akt and phosphoinositide-dependent kinases (PDK1 and PDK2). The Akt is phosphorylated by PDK1 and PDK2 at Thr 308 (solid circles) and Ser 473 (solid triangles), respectively. Class IA, II, and III PI3 kinases are expressed in HASM cells (15, 24). Our findings confirmed that p110α and -δ subunits of the class IA PI3 kinases mediate TNF-α–induced CD38 expression in HASM cells. Our previous studies showed that ERK, p38, and JNK MAP kinases mediate TNF-α–induced CD38 expression in HASM cells through transcriptional and posttranscriptional mechanisms (8). Transcriptional regulation of the CD38 gene involves activation of the transcription factors NF-κB and AP-1 (8, 36), although the PI3 kinase role in CD38 expression does not appear to be mediated through these transcription factors. The present study found that PI3 kinase does not regulate CD38 expression through modulating CD38 mRNA stability. We speculate that transcription factors other than NF-κB or AP-1 mediate the PI3 kinase effects in CD38 expression in HASM cells. In certain cell types, cross talk between PI3 kinase and MAPK signaling pathways has been reported at the level of c-Raf and Akt (26). However, findings from our laboratory and others did not support the existence of such a cross talk mechanism (17).